Resident Effect Perception in Urban Spaces to Inform Urban Design Strategies

Abstract

In the field of urban design, current research has shifted towards resident preference perception and computer-aided design methods that rely on deep learning techniques. In this study, we aimed to provide a quantitative design method for urban space design that could take into account the preferences of different populations. Through empirical research, we collected real urban space and population data, which we then quantified using advanced intelligent recognition tools based on deep learning techniques. Our ensuing analysis illuminated the intricate interplay between constituent elements of urban spaces and the structural and emotional changes of residents. By taking into account the specific driving relationships between each element and residents, we proposed a new evaluation methodology for constructing an intelligent design evaluation model for urban spaces. This intelligent design evaluation model was subsequently used to evaluate the urban space both pre- and post-design. The standard deviation of the difference results demonstrated that the design option (SD value = 0.103) and the desired option for Space 1 were lower than the current option (SD value = 0.129) and the expected scheme. Our findings provide quantitative configuration strategies and program evaluation for urban space design, thus helping designers to design urban spaces that are more popular with residents.